Fusion of multi-source satellite data and DEMs to create a new glacier inventory for Novaya Zemlya

Abstract

Monitoring glacier changes in remote Arctic regions are strongly facilitated by satellite data. This is especially true for the Russian Arctic where recently increased optical and SAR satellite imagery (Landsat 8 OLI, Sentinel 1/2) and digital elevation models (TanDEM-X, ArcticDEM) are becoming available. These datasets offer new possibilities to create high-quality glacier inventories. Here, we present a new glacier inventory derived from a fusion of multi-source satellite data for Novaya Zemlya in the Russian Arctic. We mainly used Landsat 8 OLI data to automatically map glaciers with the band ratio method. Missing debris-covered glacier parts and misclassified lakes were manually corrected. Whereas perennial snow fields were a major obstacle in glacier identification, seasonal snow was identified and removed using Landsat 5 TM scenes from the year 1998. Drainage basins were derived semi-automatically using the ArcticDEM (gap-filled by the ASTER GDEM V2) and manually corrected using fringes from ALOS PALSAR. The new glacier inventory gives a glacierized area of 22,379 ± 246.2 km2 with 1474 glacier entities > 0.05 km2. The region is dominated by large glaciers, as 909 glaciers < 0.5 km2 (62% by number) cover only 156 ± 1.7 km2 or 0.7% of the area, whereas 49 glaciers > 100 km2 (3.3% by number) cover 18,724 ± 205.9 km2 or 84%. In total, 41 glaciers are marine terminating covering an area of 16,063.7 ± 118.8 km2. The mean elevation is 596 m for all glaciers in the study region (528 m in the northern part, 641 in the southern part). South-east (north-west) facing glaciers cover > 35% (20%) of the area. For the smaller glaciers in the southern part we calculated an area loss of 3.4% (-52.5 ± 3.1 km2) from 2001 to 2016.